ADHD is one of the most frequently occurring disorders of children and adolescents, and represents a significant public health problem. A plethora of scientific data documents the robust efficacy [dj1]and safety of psycho stimulants for the management of ADHD symptoms and associated impairments. However, there is also considerable variability in response, with many individuals reporting lower levels of tolerability and palatability. Moreover, only a subgroup of those receiving stimulant treatments achieves a response that approximates normal functioning. The recent approval of atomoxetine (ATX) offers a non-stimulant alternative for the treatment of ADHD. ATX is structurally unrelated to the stimulants and appears to work via a different mechanism. Thus, ATX may be an effective treatment for those who cannot tolerate stimulants or who have an inadequate response, and responders to the two treatments may have different clinical and neurobiological characteristics. This two-site study will evaluate the relative efficacy, tolerability and palatability of Concerta methylphenidate (MPH) and ATX in the treatment of children and adolescents with ADHD, using a randomized, double blind, crossover design. To date, no study has compared the leading stimulant and non-stimulant treatments in large numbers of youth, using the most effective doses for each treatment, and evaluating response to both treatments in the same individuals. We will test the hypothesis that MPH and ATX have similar overall efficacy in 320 children, with sufficient statistical power to detect even a small effect size difference between the treatments. We will examine "normalization" of response, as well as more traditional outcomes, because the former more closely reflects the desired endpoint of treatment. We will additionally evaluate the potential moderating effects of genotype. We hypothesize that regardless of group differences in efficacy in ADHD symptom reduction, there will be differential response and tolerability within individuals, which will be associated with distinct neurobiological characteristics. These data will be vitally important to patients, families, and clinicians in identifying those factors within individuals that predict optimal response to the two treatments, and lead to a reduction in burden of illness for youth with ADHD and their families.